Joseph Kei

Determining the risk factors for otitis media in a clinical sample of children

This is a Master of Philosophy project.

Children aged 2-5 years have the highest incidence rates of otitis media (OM).The pathogenesis of OM is multi-factorial with numerous risk factors. An appreciation of the risk factors for OM is crucial as education of the community to mitigate these factors is amongst the most effective strategies to reduce the burden of OM. Although the prevention and treatment of OM in Aboriginal and Torres Strait Islander (ATSI) children have been broadly studied, risk factors for OM in the general population of Australian children have not been systematically investigated. This project aims to identify the risk factors for OM with a view to reduce the incidence of OM in young children.

Evaluating the predictive accuracy of wideband tympanometry to detect otitis media with effusion in children

This is a PhD project.

Early detection of otitis media (OM) is vital for timely management of the disease. However, diagnosing OM in young children (2-5 years) presents a great challenge to audiologists and otolaryngologists. Traditional audiologic measures such as pure tone audiometry, tympanometry and otoacoustic emission tests are failing to diagnose otitis media with effusion (OME) in young children because these measures are not highly accurate and young children are not always cooperative in undertaking these tests, making the diagnosis of OME unreliable or uncertain. This issue can be addressed by using an innovative wideband tympanometry (WBT) test that is capable of not only showing abnormal middle ear function results due to OME, but also determining if a significant hearing loss is present. This project aims to determine the predictive accuracy of WBT in diagnosing OME in young children using a statistical modelling approach and machine learning algorithms.

Automatic detection of otitis media in otoscopic images obtained from children using deep metric learning

This is a PhD project.

Otitis media (OM), inflammation of the middle ear, affects 80% of Australian children overall by the age of 3 years. Clinically, OM can be broadly classified into acute otitis media (AOM) and otitis media with effusion (OME). AOM is the second most common cause for a visit to the medical doctor. With bacterial infection, the onset of AOM is usually abrupt, with fever and irritability being the common symptoms. Children with AOM is usually treated with antibiotics, and it is the single diagnosis responsible for most prescriptions of antibiotics. OME is the most common cause of acquired hearing loss in childhood with 80% of all children under 4 years old have had at least one episode of the disease. OME is insidious with no obvious infection symptoms apart from a mild to moderate hearing loss.

Otitis media is mostly diagnosed with the use of an otoscope, allowing the medical doctor to get a visual image of the tympanic membrane. Ear-Nose-Throat specialists (ENTs) usually use more advanced and specialized otoscopes to make a diagnosis. Modern otoscopes are equipped with digital cameras, making the images suited for automated detection of OM. Research has indicated that paediatricians correctly distinguished among No-OM, OME, and AOM 50% of the time, while the accuracy of the ENTs was 75%. These results indicate the need for ENTs or properly trained primary care physicians to better diagnose otitis media. The aim of this research is to evaluate the predictive accuracy of classifying No-OM, OME, and AOM in otoscopic images obtained from children using a machine learning (artificial intelligence) approach.

Investigating the feasibility of providing children with chronic otitis media with cartilage-conduction hearing aids

This is a PhD project.

Hearing Australia provides hearing aids for children with hearing loss. While appropriate amplifications are needed for children with chronic otitis media (OM) to reduce the effects of auditory deprivation on speech, language, psychological and cognitive development, the uptake and low use of hearing aids by these children presents a challenge to audiologists and professionals working with their families. As these children may experience chronic OM and otorrhoea (discharging ear), they are prescribed with a bone-conduction hearing aid (BCHA), whereby a transducer (bone vibrator) has to be held tightly against the mastoid bone using a headband. However, there are disadvantages in using the BCHAs resulting in low uptake and use of these aids. To overcome the shortcomings of BCHAs, Professor Hosoi of Nara Medical School in Japan developed a new behind-the-ear cartilage-conduction (CCHA) for use with people with chronic outer and middle ear conditions. As the CCHA does not occlude the ear canal, it would be suitable for children with chronic OM or discharging ears. This project aims to evaluate the functional benefits of wearing a CCHA versus a BCHA in terms of hearing capability and usage by children with chronic OM.